Lead Time vs. Cycle Time: The Metrics That Matter
Lead Time and Cycle Time are two of the most powerful metrics for understanding and improving a team’s delivery flow. Although they sound similar, they measure different things.
Definitions
Lead Time
From request to delivery. Starts when an item is created in the backlog and ends when it reaches the customer.
“The customer requested it on the 1st, it was delivered on the 15th. Lead Time: 14 days.”
Cycle Time
From work start to delivery. Starts when the team actually begins development and ends when it’s done.
“Development started on the 5th, finished on the 15th. Cycle Time: 10 days.”
The Difference in Practice
[Backlog] ---( waiting )---> [Dev starts] ---( working )---> [Delivered]
|--------- Lead Time ----------|
|--- Cycle Time ---|
Lead Time will always be greater than or equal to Cycle Time. The difference between the two is the waiting time in the backlog.
What Each Metric Reveals
High Lead Time
- Items sit in the backlog for too long before being worked on
- May indicate a bloated backlog or inefficient prioritization
- Customers notice: “they took forever to deliver something simple”
High Cycle Time
- The work itself is slow
- May indicate complexity, technical bottlenecks, or excessive WIP
- The team notices: “it takes us too long to implement”
Percentiles Over Averages
Don’t use averages — they hide important outliers.
Use the 85th or 90th percentile: “85% of our items are delivered within 7 days.” That’s far more honest and useful than saying “average of 4 days” (which ignores the items that took 15 days).
Smarter Goals with Metrics
Reducing Lead Time
- Prioritize more frequently
- Keep the active backlog lean
- Use T-shirt sizing to identify quick wins
Reducing Cycle Time
- Limit WIP
- Invest in CI/CD automation
- Sharpen your Definition of Done to avoid rework
- Speed up code reviews
Industry Benchmarks
High-performing teams typically achieve:
- Lead Time for changes: less than 1 day (DORA metric)
- Cycle Time: 1-3 days for small tasks
- Deploy failure rate: less than 5%
If your numbers are significantly higher, there are clear opportunities for improvement.
Relationship with Estimation
Historical Cycle Time can complement estimates:
“Similar tasks to this one (CRUD with validation) averaged 3 days of Cycle Time over the last 2 sprints. We can use that as a reference.”
This is especially useful in Kanban, where there are no sprints to calibrate velocity against.
Conclusion
Lead Time and Cycle Time tell different stories about the same process. Use both to diagnose where the bottleneck lies — in the wait (backlog) or in the execution (development). Reducing both is the path to faster, more predictable deliveries.